Claude docs changes for June 9th, 2026 [diff]

Executive Summary

Claude Code 2.1.169 adds --safe-mode flag to disable all customizations for troubleshooting, and a /cd command to change working directory without breaking prompt cache
New disableBundledSkills setting lets organizations hide bundled skills and built-in slash commands from the model
Fixed enterprise managed MCP policies not being enforced on reconnect and during first session after install, plus slow cold starts for orgs without remote settings
API docs clarified that Skills work with all code execution tool versions (code_execution_20260120 and later), not just the original code_execution_20250825
Advisor tool documentation now notes it accepts generic tool properties: cache_control, allowed_callers, defer_loading, and strict

Added --safe-mode flag (and CLAUDE_CODE_SAFE_MODE env var) to start Claude Code with all customizations disabled (CLAUDE.md, plugins, skills, hooks, MCP servers) for troubleshooting
Added /cd command to move a session to a new working directory without breaking the prompt cache mid-session
Added disableBundledSkills setting and CLAUDE_CODE_DISABLE_BUNDLED_SKILLS env var to hide bundled skills, workflows, and built-in slash commands from the model

/workflows now opens immediately even while a turn is in progress
Improved TaskCreate reliability: malformed inputs are repaired automatically and validation errors for unloaded tools include the schema
Reduced CPU usage while responses stream and during spinner animations
Restored a default 5-minute idle timeout on Vertex/Foundry so a stalled stream aborts instead of hanging indefinitely; set API_FORCE_IDLE_TIMEOUT=0 to opt out
Remote-managed settings with an invalid entry now apply their remaining valid policies and surface the validation error instead of silently dropping the whole payload
Background sessions now preserve --ide, --chrome, --bare, --remote-control, and other flags across retire→wake
The "CLAUDE.md is too long" warning threshold now scales with the model's context window
Improved error message shown when an organization has disabled API key authentication

Fixed enterprise managed MCP policies (allowedMcpServers/deniedMcpServers) not being enforced on reconnect, IDE-typed configs, --mcp-config servers during first session after install, or before remote settings loaded; also fixed slow cold starts for orgs without remote settings
Fixed a ~30-50ms UI stall at the start of each turn for macOS users logged in with claude.ai credentials
Fixed claude -p being slow or appearing to hang on Windows while waiting for slash-command/skill scan (regression in 2.1.161)
Fixed claude agents --json omitting blocked and just-dispatched background sessions; added --all to include completed sessions, plus new id and state fields
Fixed background agents ignoring project-level settings env values (e.g. ANTHROPIC_MODEL) when dispatched onto a pre-warmed worker
Fixed stale permission and dialog prompts reappearing every time you reattached to a remote session whose worker had died while waiting on them
Fixed footer hints (e.g. "esc to interrupt") not showing for users with a custom statusline
Fixed Remote Control getting stuck on "reconnecting" after resuming a session when an OAuth token refresh happened simultaneously
Fixed Git Credential Manager's "Connect to GitHub" popup appearing on Windows at startup when background git commands ran without cached credentials
Fixed untrusted project settings being able to set OTEL client-certificate paths without trust confirmation

No significant changes.

Clarified that Skills work with all code execution tool versions: the quickstart examples use code_execution_20250825, but Skills also work with newer revisions (code_execution_20260120 and later) — any version satisfies the Skills requirement as long as its tool type and beta header are kept consistent. [line 79] [Source]

Added documentation noting that the advisor tool accepts the generic tool properties available on any tool definition: cache_control, allowed_callers, defer_loading, and strict, with a reference to the Tool reference for their semantics. [line 101] [Source]